Recorded shoplifting fell 4% in England and Wales in the year ending March 2026, to 507,086 offences. Over the same period, recorded business robbery rose 91%, to 29,669 offences. For retail security teams, those headlines raise an obvious question: what do they tell us about safety on the shop floor?
Both figures come from the ONS bulletin for the year ending March 2026. A clarification of police recording rules in April 2025 contributed to the business-robbery increase. The bulletin does not quantify that contribution, making it essential to understand the recording issue before interpreting the rise.
The figures offer useful context for officers and managers dealing with theft, threats and violence. Assessing whether a particular store has become safer requires a closer look at its own incidents.
What is the difference between shoplifting and business robbery?
The main difference is force or threats used to steal. Shoplifting is theft from a shop. Robbery involves stealing while using force, or putting or seeking to put someone in fear of immediate force, in order to steal.
The legal offence is robbery; “robbery of business property” is the statistical category distinguishing it from robbery of personal property.
| Shoplifting | Business robbery | |
|---|---|---|
| Typical example | Someone conceals goods and leaves without paying, without using force or threats. | Someone threatens a cashier and steals money from the till. |
| Force or threats | Not required for theft. | Must be connected to the stealing. |
| Weapons or injuries | Not required. | Neither is necessary: threats can be enough. |
| Value stolen | Does not determine whether the incident is robbery. | Robbery can involve low-value goods. |
The force or threat can be directed at a shopworker, security officer or another person. It does not have to be directed at the owner of the stolen property.
If a shoplifter becomes violent when challenged, the incident can amount to robbery, but this is not automatic. Crown Prosecution Service guidance on robbery explains that the force or threat must occur immediately before or during the theft and be used to carry it out.
Courts recognise that stealing can be a continuing act, so violence while leaving with the goods may qualify. An unrelated assault after a completed theft is not automatically robbery. The circumstances and sequence matter.
Why the business-robbery recording clarification matters
According to ONS’s explanation of the clarification, forces were reminded that robbery of business property should be recorded where someone steals from retail premises and uses or threatens violence against a shopworker, security guard or another person. The clarification also covers violence or threats used to escape.
This clarified how incidents should be recorded; it did not introduce a new offence. ONS says it probably also had a small effect in reducing recorded shoplifting. The two trends therefore need to be considered together.
The bulletin does not separate the business-robbery increase into changes in actual offending and changes in recording. Nor does the 91% rise represent a count of attacks on security officers: business robbery includes incidents involving other victims.
For an officer writing an incident report, the practical point is to capture the full sequence. If someone steals goods and threatens a member of staff during the incident, the report should describe the threat, who it was directed at and when it happened, alongside the property loss. The value of the stolen item alone cannot describe the harm or risk involved.
Shoplifting growth has slowed, but recorded offences remain above half a million
The latest decline follows a series of releases in which annual shoplifting growth slowed before turning negative:
| Year ending | Recorded shoplifting offences | Annual change |
|---|---|---|
| December 2024 | 516,971 | +20% |
| March 2025 | 530,324* | +20% |
| June 2025 | 529,994 | +13% |
| September 2025 | 519,381 | +5% |
| December 2025 | 509,566 | −1% |
| March 2026 | 507,086 | −4% |
These are overlapping rolling 12-month periods, with adjacent periods sharing nine months. March 2025 uses the revised count of 530,324 published in the March 2026 bulletin; the original March 2025 release reported 530,643. Other counts and the rounded annual changes reproduce their respective releases. This is a comparison of published bulletins, rather than a fully revised historical series.
The published figures show strong annual growth giving way to a decline. However, they cannot demonstrate an equivalent reduction in all theft experienced by retailers.
ONS says shoplifting reporting rates are very low. Recorded totals can be affected by the number of incidents, willingness to report them, police recording practices and police activity. The 4% fall does not establish how much each factor contributed.
Earlier reporting periods also showed business-robbery increases
The published series includes substantial increases for periods ending before April 2025:
- Year ending December 2024: business robbery rose 54%, to 14,707 offences.
- Year ending March 2025: business robbery rose 50%, to 15,520 offences.
These earlier periods provide useful context, but their timing does not isolate the cause of the increases. Both bulletins were published after April 2025, and they do not establish whether later recording decisions affected incidents from earlier periods. Reporting and recording changes also need to be considered alongside possible changes in offending.
Personal robbery presents a different picture. Recorded offences fell 9% to 57,693 in the year ending March 2026, while recorded business robbery increased. The overall robbery total conceals this distinction, but the contrast does not establish that offenders switched from targeting individuals to businesses. Differences in reporting and recording could also affect the categories.
Do falling theft figures mean a safer shift?
The national figures cannot answer that question for an individual store or security contract. They do not establish:
- Whether threats or assaults against security officers specifically have increased or decreased.
- Whether repeat offenders account for a growing share of serious incidents.
- Whether particular stores have become safer for staff and customers.
- Whether staffing, equipment or local policing changes caused an improvement.
Managers need to distinguish stock losses, incident frequency and harm to people. A store could experience fewer thefts while recording the same number, or more, of threats and assaults. That is a possibility to test against local records, rather than a finding from these ONS figures.
The management decisions behind staff protection are explored in Tom Richmond’s opinion piece on retail workplace violence and investment in guarding. It offers a separate perspective on retailers’ priorities, rather than evidence explaining these statistical changes.
What consistent incident reporting can reveal
Local records can help assess site-level risk when reporting practices are consistent. Security teams reviewing incidents should distinguish:
- Theft: whether attempted or completed, the goods involved and the circumstances.
- Threats: what was said or done, who was targeted and when it happened.
- Assaults and injuries: details recorded distinctly within the incident record, alongside the property loss.
- Evidence: available CCTV, body-worn video and witness accounts.
- Police reporting and classification: whether reported, the reference number and the classification applied, where known.
- Patterns: repeat incidents, locations, times and evidenced links between reports.
Avoid counting linked categories as separate incidents. In a site’s incident log, one confrontation involving theft, threats and assault can be recorded as one incident with each element identified. This is a way to organise internal records, not an instruction on how police should count offences.
Missing or unusable footage can weaken the evidence available to explain what happened. Equipment and evidence-handling arrangements therefore deserve attention alongside the reporting form. TPSO has explored related equipment concerns in Cut-Price Body Cameras Are Undermining Security Standards.
Local records have limitations too. Incidents may go unreported, and a new reporting system or greater encouragement to report threats can raise recorded totals without an equivalent rise in offending.
Before comparing one year with another, managers should ask:
- Have thefts, threats and assaults followed the same pattern?
- Have definitions, reporting systems or reporting rates changed?
- Have operating hours, staffing or the number of sites covered changed?
- Have police classifications or responses changed?
- Do officers’ accounts reveal incidents or patterns missing from the records?
Discussing those questions with officers and supervisors helps put the numbers in context. TPSO’s article on bringing risk management closer to frontline practice explores the value of clear communication and officer involvement in risk discussions. A local example can explain one team’s experience, but should not be presented as proof of a national trend.
Measure safety as well as theft
The fall in recorded shoplifting follows a period of sustained increases and is welcome news. The rise in recorded business robbery also deserves attention, with the April 2025 clarification central to its interpretation.
For security leaders, the next step is to examine threats and harm alongside theft totals. Consistent reporting, usable evidence and frontline input provide a stronger basis for assessing staff safety than a national headline alone.
Are theft, threats and assaults moving in the same direction at your sites, and do your incident reports distinguish between them?
Source note: This analysis draws on six ONS Crime in England and Wales bulletins, covering overlapping annual periods ending December 2024, March 2025, June 2025, September 2025, December 2025 and March 2026. The latest annual changes are those reported in the March 2026 bulletin, using its comparators. Historical figures retain the versions described in the table note and text. All offence totals discussed here are police-recorded figures.